Reflection language model
Beam
the first open-weights model from Reflection, a 501B-total, 23B-active mixture-of-experts built for agentic tool use, coding and general reasoning. it reads text only and its weights ship under Apache 2.0.
01
price
input tokens
output tokens
cached input
per 1M tokens. octomind bills the listed provider price, with no markup, on every plan, including the free one (pay as you go from credits).
02
what real work costs
| workload | cost |
|---|---|
| one chat reply 3,000 tokens in, 500 out | $0.00058 |
| summarise a 100-page report 50,000 tokens in, 1,500 out, one request | $0.0069 |
| 1,000 classification calls 500 tokens in, 20 out, per call | $0.07 |
| coding-agent session 2M tokens in across ~60 turns, 60,000 out, no cache | $0.27 |
| same session, cached 85% of input read from the prompt cache, the rest written to it | $0.09 |
at the listed rate. token counts are typical sizes, not measurements.
03
specs
- provider
- Reflection
- context window
- 1M tokens — ≈ 1,500 pages of prose or 100,000 lines of code
- max output
- 128k tokens
- input
- text
- output
- text
- structured output
- yes — JSON schema
- reasoning mode
- yes
- prompt caching
- yes — cache hits at $0.015/M
- weights
- open — Apache 2.0
- released
- October 2026
- knowledge cutoff
- June 2026
- status
- live
- model id
beam
integration notes
- Beam always reasons: Reflection's API takes low, medium, high, xhigh or max, and answers 400 to none or minimal.
- Reflection's API requires a non-empty reasoning_content on every assistant message and answers 400 without it.
- Reflection recommends temperature 0.7 and top_p 0.99; top_k, min_p and repetition_penalty can't be set.
04
benchmarks
as published by the maker; not measured by octomind.
05
when to pick it
pick it if
- you want a long-window text model priced with the flash tier
- you want Apache 2.0 weights, with BF16, FP8 and NVFP4 builds to self-host
- you run tool-calling agents on text and repeat long prompts, so cached input matters
skip it if
- you need image, audio or video input; Beam reads text only
- long autonomous coding is the job; its reported DeepSWE v1.1 score sits below GLM-5.3 Flash and DeepSeek V4.1 Flash
06
where it sits on price
blended price $0.194/M at 3 input tokens per output token: 4 live language models cost less, 39 cost more.
- Mercury 3 $0.163/M
- MiMo V2.6 Flash $0.175/M
- Beam $0.194/M
- Claude Haiku 5.5 $0.2/M
- Gemma 4 31B $0.2/M
07
use Beam in octomind
one octohub login reaches every model in the catalog. pick this one per run, per session, or as your default.
octomind login
octomind run -m octohub:beam/model octohub:beam[model]
name = "octohub:beam"08
compare Beam
- Beam vs GLM-5.3 Flash
- Beam vs MiniMax M3.1 Flash
- Beam vs Qwen3.8 Omni Flash
- Beam vs MiMo V2.6 Flash
- Beam vs Mercury 3
- Beam vs GPT-6 Luna
- Beam vs Claude Haiku 5.5
- Beam vs Muse Spark 1.3 Contributor
- Beam vs DeepSeek V4.1 Flash
- Beam vs MiniMax M3
- Beam vs MiMo V2.6 Pro
- Beam vs DeepSeek V4 Pro
- Beam vs GLM-5.3
- Beam vs Kimi K3
- Inkling 2 Small vs Beam
09
similar models
nearest live models by listed input price.
- MiMo V2.6 Flash $0.14/M input
- Claude Haiku 5.5 $0.1/M input
- GPT-6 Luna $0.1/M input
- Mercury 3 $0.1/M input
10
Beam faq
how much does Beam cost?
Beam lists at $0.125/M input, $0.4/M output, $0.015/M cached input per 1M tokens. a typical coding-agent session (2M tokens in, 60,000 out) comes to $0.27, or $0.09 when 85% of the input is read from the prompt cache. octomind bills the listed provider price, with no markup, on every plan, including the free one (pay as you go from credits).
what is the context window of Beam?
1M tokens, ≈ 1,500 pages of prose or 100,000 lines of code. a single reply can run to 128k tokens.
does Beam accept images?
no. Beam is text-only; pick a vision model from the Reflection lineup or the catalog for screenshots and diagrams.
is Beam open source?
its weights are open under Apache 2.0, so it can also be self-hosted.
when was Beam released?
October 2026, by Reflection.
is Beam free?
no. Beam bills per token at $0.125/M input, $0.4/M output, $0.015/M cached input per 1M. octomind's free plan reaches it pay as you go from credits, so you pay only for the tokens you use, at the provider price. credits carry a 3% top-up fee.
how do i use Beam in octomind?
run octomind login once, then octomind run -m octohub:beam. inside a running session, /model octohub:beam switches to it without losing context.
use Beam in octomind
choose a model when you start, then switch mid-session when the work changes.